[jneutron]

Please be forwarned this is a VERY long detailed thread.

Not a problem.

Back to some more questions about the ABX.. Peter "the Axe" at The Audio Critic magazine

Isn't he the guy with the ten fallacy thing?  7 of them are not well thought out.

Finally it is time to give readers here who seem bent on throwing out links, papers toward "scientific" knowledge that meets "their criterea" for proof of cable differences cannot exist. In the December 1995 Stereophile issue Ben Duncan provided a 10 page overview of wire measurments comparing 8 different cables and showed ABSOLUTE CONCLUSIVE MEASUREMENTS WITH GRAPHS AND EASY TO SEE DIFFERENCES! Yes you read that right, It's been almost 12 years since something scientific was presented that showed without doubt something actually measurable. It was much more than just a meager frequency response comparitive. These tests showed varying degrees of energy storage and release after the cables were hit with a tone burst.

Waveforms do not show audibility.  Audibility studies are for that.  Unfortunately, audibility studies are not sensitive to the level of localization humans are capable of.

Not enough science to make you a believer yet "Mr. Science"? In the October '95 Stereophile there was an overview from Professor Macolm Omar Hawkford using Maxwell's equations developing mathematical models describing the behavior of cables. Again, this was done over a decade ago people. The science has also vidicated the "ears".

I would not recommend hanging your hat on that article.  It contains very large errors in assumption.  Those errors completely invalidate the premise and conclusion of the article.  Had that article been subjected to peer review, and had I been one of the referees, it would have been soundly rejected because of those errors.

BTW, Malcolm is a very nice guy.  And within his discipline, he is an absolute monster..I like him.

Cheers, John

[Night_Train]

I've used heavy duty auto jumper cables with excellent results. The huge ones! Also, bicycle lock cable works too. The coiled stuff is just too hard to work with though.

[Daygloworange]

Waveforms do not show audibility.  Audibility studies are for that.  Unfortunately, audibility studies are not sensitive to the level of localization humans are capable of.

John,

Interesting. Could you elaborate a little more on that?

Cheers

[jneutron]

Waveforms do not show audibility.  Audibility studies are for that.  Unfortunately, audibility studies are not sensitive to the level of localization humans are capable of.

John,

Interesting. Could you elaborate a little more on that?

Cheers

Sure.  When you mixdown, you use the pan to move an image off axis, but no time delay.

We are sensitive to both...

To correctly position an image ten feet away... 20 inches off axis, requires a relative interchannel time delay of about 80 uSec and .05 dB relative. (note, the assumption used for the math is a spherical wavefront from a point source..).  Since you do not have such control over time, you use level to achieve the goal.  That creates a reproduction system that has sensitivity to the reproduction driver positions, hearing sensitivity, and integrated hearing exposure (we acclimate to differing localization cues in real time, this ability is ignored within the dbt test regimens.)

It would be better to setup test regimens which look for a relative shift in image using accurate localization parameters.  Detection of location of one image relative to another is far stronger since it eliminates head positioning (no head vice).

I have some graphs to depict this, but alas, this forum does not seem to support jpegs.

Cheers, John

[Daygloworange]

As a big fan of stereo miking techniques, I always allude to that when I'm talking about "soundstage" and depth and so on. Time arrival issues between channels, early and late reflections, phase shift, comb filtering and so on.

I also feel that the effects of cables and such could be measured with a proper test setup using (multiple) mic set ups taking time arrival plots from a test system's speakers. You could then do actually overlays and "see" what the differences were. I discussed it in a thread a while back, and described what I thought might be able to actually be able to be a comprehensive test for it.

I believe there is a fellow that has a software package that he made available for download here on AC that could possibly be rigged up to do such a test.

I have some graphs to depict this, but alas, this forum does not seem to support jpegs.

You have to upload them here in your Gallery first, then during your post, link to where in your gallery it's located.

There are threads here explaining how to post pictures on AC.

I'd love to see the graphs you have.

Cheers

[jneutron]

I also feel that the effects of cables and such could be measured with a proper test setup using (multiple) mic set ups taking time arrival plots from a test system's speakers. You could then do actually overlays and "see" what the differences were. I discussed it in a thread a while back, and described what I thought might be able to actually be able to be a comprehensive test for it.

I believe there is a fellow that has a software package that he made available for download here on AC that could possibly be rigged up to do such a test.

Using mikes is a problem.  Each mike picks up the other channel's information, and from that point on, the signal processing is incapable of discerning images.  I have yet to see the algorithms which are capable of doing what we already do naturally.

I have some graphs to depict this, but alas, this forum does not seem to support jpegs.

You have to upload them here in your Gallery first, then during your post, link to where in your gallery it's located.

There are threads here explaining how to post pictures on AC.

I'd love to see the graphs you have.

Cheers

Got it, thanks..now the members hafta pay the price... .  First I just hadta load magfield and calculation pics...it's the geek in me...

I was pleased to see that I can upload 2 meg files, but unfortunately, I compressed all of them for other forums which had a 100 K limit. 

I won't be limited to that anymore.

Cheers, John

[Daygloworange]

Using mikes is a problem.  Each mike picks up the other channel's information, and from that point on, the signal processing is incapable of discerning images.  I have yet to see the algorithms which are capable of doing what we already do naturally.

I don't know for sure John, as I don't use calibration mics, but it seems to me with that with proper mics (uni directional, side cancelling mics) and/or binaural set-ups, you could get at least close enough to do some pretty revealing tests on some of the important time and spatial cues that are embedded in 2 channel audio and come out of audio speakers, that could prove empirically the differences in what we (report) hear and seemingly haven't yet been able to measure between pieces of equipment and/or cables and so forth.

Cheers

[AdamM]

You could try the Blumlein pair microphone technique for this.

It's amazing for capturing spacial positioning and detail.  Many famous classical recordings are done with this technique alone - just two microphones for the whole mix!

With decent sensitive ribbon mics, you might find it's one of the best methods for attempting to capture the spacial detail

http://en.wikipedia.org/wiki/Blumlein_Pair

[Daygloworange]

You could try the Blumlein pair microphone technique for this.

You wouldn't want to use figure 8 pattern mics for something like this. You would want heavy rearward and side rejection pickup patterns. Ribbon mics don't typically have the crisp transient response of condenser mics, and typically exhibit a rolled of high frequency response.

Cheers

[Vapor Audio]

Hey, I'm proud to say that I recently passed a blind test at bpape's house.  After about 2 hours of listening to the Sentinels Bryan jumped behind his rack and said "check this out".  I had no idea at all what he changed, my initial thought was that he was making a change that he though should yield an improvement.  But as soon as the music came back I got a sour look on my face and told him the entire stage just shrunk along with dynamics. 

I was quite happy with myself when he told me he swapped in some cheaper Van Den Hul (I think) interconnect, and that he was expecting a subjectively worse sound.  Of course his room and system are totally capable of resolving minute changes, and in thise case they were quite obvious ...

[AdamM]

Wouldn't the room amplify and help spread any differences at the source?

Sure, it would be coloured, but you aren't going for the pure signal, you're going for the delta between two signals.  Therefore, if the room acts in a way to accentuate those nuances - 'stretching out' the minute differences - then you might be better off going for that stuff.

It's like you're using the room as an amp.  So what if the room resonates differently at different frequencies, wouldn't that too increase the delta between the samples and therefore help show differences?

Yeah, condensers are better for detail in the upper registers...

It might be and interesting side experiment regardless

/A

[Daygloworange]

Wouldn't the room amplify and help spread any differences at the source?

That's the problem. Room modes, room gain, comb filtering and room reflections would smear the sample.

Cheers

[AdamM]

That's the problem. Room modes, room gain, and reflections would smear the sample.

I know, but aren't we comparing?

We would use the room as an amp, as a way to amplify the differences between the samples.

It's like, you don't plug your speakers in to the record player directly to hear differences between cartridges, you amp the signal to increase the signal - good and bad.

Comparing the cables recorded in a big resonant room would amplify any minute differences between them.  Doing a diff between the two signals - crap, resonances, room modes and all, would reveal more of the delta than the pure differences alone.

/A

[Daygloworange]

Nah. Room bad.

Seriously though. Think about it. The room effects are noise. 

We are talking about a controlled experiment. We want to isolate what is coming directly out of the speakers. Anything else is unwanted. It will have a negative effect on the time/domain spatial cues that we are trying to isolate and capture.

It's like, you don't plug your speakers in to the record player directly to hear differences between cartridges, you amp the signal to increase the signal - good and bad.

Yeah, but you wouldn't hear the differences as well if your amp had poor frequency response, and high THD. Right?


Cheers

[AdamM]

Maybe i'm not getting something.

You're trying to determine differences between the signals.

Therefore, the only valid data is the delta, not the signal alone.

Using a room, with all it's associated crap, would grab onto key frequencies and make a mess of things - but that's good.

Because the room is a constant between both tests, you throw it out of the equation , and you're left with this lovely data which will give a stronger diff, than just the pure signals alone.

It's a common scientific technique to amplify the signal in whatever arbitrary way - as long as that's constant between both experiments - so you're left with more information to mine and use to determine a difference.

Like enlarging a photo.  Noisy and inaccurate, but consistent.    My suggestion is to look at an 11x14 enlargement of the problem, not a postage stamp sized one.

You are doing a comparison, no?  The delta is the information you want right?

/A

[Daygloworange]

No. The room has a formant. It has a character that it imparts of it's own. That's not what we are trying to snapshot here. We want the room out of the equation.

That's why speaker designers without huge anechoic chambers shoot gated responses and from close distances.

Think about it. 

Cheers

[Daygloworange]

To use a camera analogy Adam, say you wanted to compare 2 different brands of film for color, sharpness and so on, would you shoot them both through a camera with a dirty lens that has a filter on it and and distorted optics, and then view them on an uncalibrated computer monitor?

You're shooting the 2 rolls with the same lens, but the lens is adding elements that make whatever differences between the 2 films buried in the noise.

Cheers

[AdamM]

I am thinking about it

If you are recording for the pure performance alone, indeed, the room is a horrible idea.

However, i think you're missing that, in the case of comparing two cables, using a system which consistently amplifies the base signal is desirable in this instance.

The speakers impart distortion and colouration, yet you're able to throw them out of the equation because they're constant. So why not the room?

Sound like crap?  sure!  But that's not what this is about, this is about measuring differences.  It's another very desirable way to amplify what will be extremely small differences out of the ends of those wires.  The speakers do their part - you're happy for that - and a room too can also assist in the process, considering it's constant, and if the only information you seek is the difference between cables.

Subtle time shifts, frequency response, all that would be amplified by the room.  The speaker off-axis differences between the signals would also be captured in this way, and not in the method you're proposing.   

It's like doing a diff on two different convolution reverb samples, if the room is the same, you go part way to detecting differences in the initial signal.

Anyways, just an unconventional DIY idea to compare cables.  It holds ground technically.

/A

.

[AdamM]

would you shoot them both through a camera with a dirty lens that has a filter on it and and distorted optics, and then view them on an uncalibrated computer monitor?

Yes i would. Anyone would, if they realized it added another stage of amplification which was constant, and therefore could be thrown out.  It would continue to help amplify the subtle differences in the signals.

Wire end [what we want to compare]  Amp-> speaker-> ear.    Two stages of amplification

Wire end -> Amp -> speaker -> room->  -!all constants!-  Three stages of amplification

/A

[AdamM]

I know, this last little back and forth has been a little tongue-in-cheek

I agree that it's a very far from ideal test scenario.  However as a tool in your repertoire, i feel it does have merit.

Introducing noise into a test scenario is a bit crazy.  We'd ideally simplify the entire thing to best determine cause and effect.

However, detecting differences between cables is an interesting thing.  We're talking extremely subtle differences!

Assuming a test rationale of:

• VERY subtle differences
• The fact there IS a difference is more important than what that difference is
• Something we can do cheaply in a home setting with non-specialist gear

My 'noisy room' suggestion will do very little to help determinewhat the differences are, just that there may be.  Adding noise to this test is like adding a function to a function. Arg!

I just think that the differences are going to be sooooooo subtle, that a noisy test which just shows 'how much' but not 'what' might be a valid additional test metric.

Cheers,
/A